v0.7.16

What's Changed 🚀

✨ Features

• feat(droid): Add daft.datasets.droid @srilman (#7089)
• feat(temporal): add Spark-style timezone conversions @BABTUNA (#6919)
• feat: add top-level daft.concat() for concatenating multiple DataFrames @Liusixuuu (#7105)
• feat(file): Add a file_exists expression @srilman (#7140)
• feat: Add DataFrame.to_torch_dataloader @srilman (#6997)
• feat(daft-ext): add scalar #[daft_func] proc macro @rchowell (#7043)
• feat(iceberg): support SQL read_iceberg ignore_corrupt_files @jackylee-ch (#7130)
• feat: add Spark-compatible string functions (translate, substring_index, soundex, ascii, chr, space) @XuQianJin-Stars (#7070)
• feat(iceberg): support branch and tag reads in SQL @jackylee-ch (#7084)
• feat: add support for range partition hints in clustering spec @euanlimzx (#7050)
• feat: tests for aligned asof join @euanlimzx (#7072)
• feat: add ignore_corrupt_files option to read_parquet, read_csv and read_iceberg @chenghuichen (#6520)
• feat(ray): Implement dynamic scale-in for RaySwordfishActor @huleilei (#5903)
• feat(delta-lake): support column mapping for reads @aaron-ang (#7005)
• feat(functions): add string distance/similarity functions @nish2292 (#7068)

🐛 Bug Fixes

• fix(scan): estimate scan size from materialized buffer, not encoded size @madvart (#7161)
• fix(scan): use parquet metadata for scan task size estimates @desmondcheongzx (#6542)
• fix(session): list_namespaces mirrors list_tables resolution @YuangGao (#7144)
• fix(tracing): honor OTEL resource env config @RitwijParmar (#7061)
• fix(parquet): Use chunk size for ArrayReader buffer size @srilman (#7149)
• fix(ci): pin maturin <1.14 to keep --timings=html working @XiaoHongbo-Hope (#7145)
• fix: Fix flaky dataloader test @srilman (#7142)
• fix: detect unschedulable actor UDFs and warn on over-requested concurrency @XiaoHongbo-Hope (#7113)
• fix(session): list_tables mirrors get_table resolution and returns Identifier @YuangGao (#7126)
• fix: actor UDF followed by BlockingSink deadlocks due to held channel sender @XiaoHongbo-Hope (#7115)
• fix(sql): allow read_parquet file options @jackylee-ch (#7128)
• fix(parquet): cast map values to explicit schema @jackylee-ch (#7093)
• fix(mcap): support http urls through fsspec handler @RitwijParmar (#7062)
• fix(deps): bump ray floor to >=2.11.0 @rohitkulshreshtha (#7116)
• fix(native-runner): assign unique plan-cache fingerprint when caller omits one @TechyMT (#7088)
• fix(flotilla): tolerate node churn during worker startup @desmondcheongzx (#7100)
• fix(deps): gate vllm to Linux so macOS/Windows resolve without CUDA wheels @rohitkulshreshtha (#7095)
• fix: pass options in Gravitino PostgreSQL read method @qingfeng-occ (#7047)

🚀 Performance

• perf(inline-agg): add BoolAnd and BoolOr accumulator types @BABTUNA (#6984)
• perf(inline-agg): add Product accumulator type @BABTUNA (#6975)

📖 Documentation

• docs: explain lazy execution in architecture flow @everettVT (#6994)
• docs: collapse sidebar nav and group Guide sections @colin-ho (#7131)
• docs: update uuid function docstring @everettVT (#7106)

✅ Tests

• test(hf): retry transient network errors in HF integration tests @XuQianJin-Stars (#7120)
• test: tolerate HuggingFace Hub HTTP 429 in classifier and IO tests @XuQianJin-Stars (#7080)
• test(parquet): cover read_parquet edge cases @jackylee-ch (#7085)

🔧 Maintenance

• chore: upgrade arrow-rs from 57.1.0 to 59.0.0 @mikedep333 (#7141)
• chore: Replace TOS support with OpenDAL @srilman (#7135)
• chore: upgrade opendal from 0.55 to 0.57.0 @XuQianJin-Stars (#7037)

👋 New Contributors

• @Liusixuuu made their first contribution in #7105
• @nish2292 made their first contribution in #7068
• @TechyMT made their first contribution in #7088
• @RitwijParmar made their first contribution in #7061
• @XiaoHongbo-Hope made their first contribution in #7113

Full Changelog: v0.7.15...v0.7.16